Method and system of generating on-demand video of interactive activities

Abstract

A method and system for generating on-demand video is disclosed. The method includes creating a plurality of video snippets from a plurality of videos comprising trainer performed activities. The method further includes generating a set of input vectors for at least one activity dimension based on a predetermined on-demand preferences. For each of the at least one activity dimension, the set of input vectors are compared with a set of activity vectors associated with each of the plurality of video snippets, and a distance between each of the set of input vectors relative to the set of activity vectors is determined. A set of video snippets is identified where the distance is below a predefined threshold. The identified set of video snippets are combined according to at least one of the predetermined on-demand preferences, and the on-demand video is generated.

What is claimed is: 
     
         1 . A method for generating on-demand video, the method comprising:
 creating, by a multimedia processing model, a plurality of video snippets from a plurality of videos comprising activities performed by a trainer;   generating, by the multimedia processing model, a set of input vectors for at least one activity dimension based on a set of predetermined on-demand preferences;   comparing for each of the at least one activity dimension, by the multimedia processing model, each of the set of input vectors with a set of activity vectors associated with each of the plurality of video snippets;   determining for each of the at least one activity dimension, by the multimedia processing model, distance of each of the set of input vectors relative to the set of activity vectors associated with each of the plurality of video snippets;   identifying, by the multimedia processing model, a set of video snippets from the plurality of video snippets, wherein the distance determined for the set of video snippets relative to the set of activity vectors associated with each of the plurality of video snippets is below a first predefined threshold;   combining, by the multimedia processing model, the set of video snippets based on at least one of the set of predetermined on-demand preferences, wherein combining the set of video snippets comprises:   determining an order of combining the set of video snippets based on predefined criteria;   selecting, for each set of consecutive video snippets in the determined order, a connective video snippet from a set of connective video snippets; wherein selecting the connective video snippet from a set of consecutive video snippet comprises:   determining a first distance of a set of end edge vectors of a preceding video snippet from the set of consecutive video snippet with a set of start edge vectors of each of the set of connecting video snippets;   determining a second distance of a set of start edge vectors of a succeeding video snippet from the set of consecutive video snippet with a set of end edge vectors of each of the set of connecting video snippets; and   selecting the connective video in response to comparing the set of end edge vectors of the preceding video snippet and comparing the set of start edge vectors of the succeeding video snippet, wherein an average of the first distance and the second distance is the lowest relative to an average distance computed for the remaining set of connective video snippets;   interleaving the connective video snippet between the associated set of consecutive video snippets; and   generating, by the multimedia processing model, an on-demand video based on the combining.   
     
     
         2 . The method of  claim 1 , wherein combining the set of video snippets comprises: 
       determining, for each of the set of video snippets, a set of start edge vectors and a set of end edge vectors;
 selecting a first video snippet from the set of video snippets; 
 comparing the set of end edge vectors of the first video snippet with the set of start edge vectors of each of the remaining set of video snippets; 
 selecting a second video snippet from the remaining set of video snippets, wherein the distance between the set of end edge vectors of the first video snippet and the set of start edge vectors of the second video snippet is below a second predefined threshold; and combining the first video snippet with the second video snippet. 
 
     
     
         3 . The method of  claim 2 , wherein the first video snippet is selected based on the set of predetermined on-demand preferences, and wherein the first video snippet is an opening video snippet in the on-demand video. 
     
     
         4 . The method of  claim 1 , wherein combining the set of video snippets comprises:
 assigning a rank to each of the set of video snippets based on proximity of the determined distance, wherein a video snippet with high proximity is ranked highest and a video snippet with least proximity is ranked the lowest; and   combining the set of video snippets based on the rank assigned to each of the set of video snippets, wherein the highest ranked video snippet is the opening video snippet in the on-demand video.   
     
     
         5 . The method of  claim 1 , wherein the set of predetermined on-demand preferences are received from a user. 
     
     
         6 . The method of  claim 1 , wherein the set of predetermined on-demand preferences are created based on a set of standard predefined user preferences. 
     
     
         7 . The method of  claim 1 , wherein the computed distance is determined for the set of video snippets relative to the set of activity vectors as at least one of a Euclidean distances. 
     
     
         8 . The method of  claim 1 , further comprising generating, via a trained Artificial Intelligence (AI) model, the set of consecutive video snippets, wherein generating comprises:
 determining a set of vectors for each of the set of video snippets;   computing an average set of vectors based on the set of vectors determined for each of the set of video snippets; and creating, via the trained Al model, the set of consecutive video snippets based on the average set of vectors.   
     
     
         9 . The method of  claim 1 , further comprising:
 rendering, via display device, the generated on-demand video to a user;   monitoring, in real-time, via an Artificial Intelligence (AI) model, activities performed by the user in response to viewing the rendered on-demand video;   synchronizing, by the AI model, pace of rendered on-demand videos with the pace of the user in response to the monitoring; and generating, by the AI model, dynamic feedback to the user in response to the user performances.   
     
     
         10 . The method of  claim 9 , wherein synchronizing the pace of the user comprises controlling speed of the user, repeating activities, stopping the activities, or the like. 
     
     
         11 . The method of  claim 9 , wherein the dynamic feedback comprises at least one of:
 amount of calories burnt, maximum count of the at least one activity performed, maximum time spent for the at least one activity during a previous activity session of the user, incorrect posture or pace of the user while performing the at least one activity, correct posture or pace of the user to perform the at least one activity, absolute activity performance proficiency of the user, relative activity performance proficiency of the user, or best time taken to perform the at least one activity.   
     
     
         12 . A system for generating on-demand videos, the system comprising:
 a processor; and   a memory communicatively coupled to the processor, wherein the memory stores processor-executable instructions, which, on execution, causes the processor to:   create a plurality of video snippets from a plurality of videos comprising activities performed by a trainer;   generate a set of input vectors for at least one activity dimension based on a set of predetermined on-demand preferences;   compare for each of the at least one activity dimension, each of the set of input vectors with a set of activity vectors associated with each of the plurality of video snippets;   determine for each of the at least one activity dimension, distance of each of the set of input vectors relative to the set of activity vectors associated with each of the plurality of video snippets;   identify a set of video snippets from the plurality of video snippets, wherein the distance determined for the set of video snippets relative to the set of activity vectors associated with each of the plurality of video snippets is below a first predefined threshold;   combine the set of video snippets based on at least one of the set of predetermined on-demand preferences, wherein combining the set of video snippets comprises:   determining an order of combining the set of video snippets based on predefined criteria;   selecting, for each set of consecutive video snippets in the determined order, a connective video snippet from a set of connective video snippets; wherein selecting the connective video snippet from a set of consecutive video snippet comprises:   determining a first distance of a set of end edge vectors of a preceding video snippet from the set of consecutive video snippet with a set of start edge vectors of each of the set of connecting video snippets;   determining a second distance of a set of start edge vectors of a succeeding video snippet from the set of consecutive video snippet with a set of end edge vectors of each of the set of connecting video snippets; and   selecting the connective video in response to comparing the set of end edge vectors of the preceding video snippet and comparing the set of start edge vectors of the succeeding video snippet, wherein an average of the first distance and the second distance is the lowest relative to an average distance computed for the remaining set of connective video snippets; interleaving the connective video snippet between the associated set of consecutive video snippets; and   generate an on-demand video based on the combining.